Development Of Pulsed Eddy Current Detector And Research Of Metal Surface Defect Recognition Algorithm

Pipeline transportation is widely used in large-scale and long-distance transportation of oil and gas resources.However,oil and gas pipelines are prone to damage and leakage,which have a huge impact on industrial production and ecological environment.Therefore,the safe and stable operation of the oil pipeline network is extremely critical.In recent years,non-destructive testing technology(NDT)for pipeline defect detection and performance evaluation has become a major research hotspot in China.As one of the methods,pulse eddy current testing technology is studied in this paper.The main work is carried out from the following aspects:Firstly,the equivalent circuit of the excitation end and the detection end of the pulse eddy current detector are analyzed,and establish the nonlinear analytical relationship model of the pulse eddy current detection,then the equation expression of the detection voltage signal at both ends of the detection coil is obtained.Secondly,use ANSYS finite element simulation software to establish the simulation model of pulse eddy current detector,and then obtain the information of magnetic field,eddy current and detection voltage of the model.Simulate the defect in the specimen and study its influence on the physical quantities mentioned above,so these are helpful to understand the principle of pulsed eddy current detection.By changing the lift-of values of the coil in the simulation model of the pulse eddy current detector,the interference effect of lift-off effect on detection signal is obtained.Which provides theoretical guidance for the design of pulse eddy current detector and experiment subsequently.Thirdly,design the basic circuit for pulse eddy current detector,including excitation signal generator circuit,power amplifier circuit and output matching circuit,which can provide the required pulse signal for the excitation coil.Use a series of experiments and data to optimize the relevant parameters of the detector,including the frequency and voltage of the excitation signal,the size of the excitation and detection coil and so on,which improves the performance of detector.Fourthly,this thesis presents a defect size recognition algorithm based on parameter fitting feature and random forest regression.The algorithm is driven by a large number of experimental data obtained by pulsed eddy current detector,so a defect size recognition model based on parameter fitting feature and random forest regression is established.Then the model is tested with the test data of standard defects on the tested metal specimens,and the results of the model are analyzed.Finally,after studying the above four aspects,the developed pulsed eddy current detector can detect the surface defects of the tested metal specimens and predict the size of the defect.